cuneiform team mailing list archive

[Dmitry Polevoy <openocr.polevoy@xxxxxxxxx>]

Sorry for my delay with this response.

Cognitive has been reorganizing last 9 month and we hope reorganization will
be finished in May 2010. At present time official open source development in
Cogntive is frozen. I can say there are developers in Cognitive who are
interested in OpenOcr development, but we have no administrative resources.
Some times I can find a little free time to help with windows specific, but
it is very small resource.

As for tables, current position in Cognitive is to open this part of
sources, but it will be time consuming and more possible at first time like
shareware.

In my oppinion changes in refactoring branch (
lp:~serge-uliss/cuneiform-linux/refactoring<https://code.launchpad.net/%7Eserge-uliss/cuneiform-linux/refactoring>)
are very massive and agressive. And without special testing instruments
ocr-quality drops are very possible. For example, after massive changes in
type names I can't localize probable place of some changes in long-time
history because of all files were changed. I can't understand benefits of
library renaming. And it is unacceptable to add source with trouble for
project lisence to source tree (for example GNU like in
src/rdib/bmpimageloader.cpp<http://bazaar.launchpad.net/%7Eserge-uliss/cuneiform-linux/refactoring/annotate/1144/src/rdib/bmpimageloader.cpp>
*).*

We should not use words "Cuneiform" or "Puma" for public releases because of
proprietary rights. OpenOcr is more right in my oppinion.


2010/1/6 Jussi Pakkanen <jpakkane@xxxxxxxxx>

> On Thu, Dec 31, 2009 at 6:52 PM, Dmitry Polevoy
> <openocr.polevoy@xxxxxxxxx> wrote:
>
> > In attach some patch for cfio to build in MS VC 2008 (I make no check for
> > other compilers).
> >
> > But I've found broken project cpage build in MS VC.
>
> Thanks. I've been thinking what to do with this for a while now. The
> basic problem with the refactoring branch is that we get further and
> further from the original code, merging new stuff becomes more and
> more difficult. One such example is the table recognition code that
> was discussed years ago. Will Cognitive release it? Can we get it
> working?
>
> Another problem is that the rf-branch is very Windows-hostile and
> makes some architectural changes that I don't feel totally comfortable
> with. The merging also takes a lot of time because of missing files,
> commits that have been done without testing and so on. As an example
> the newest merge has been waiting a long time because I get
> segfaults/double frees and tracking them is extremely time consuming.
>
> I have even been thinking of removing the latest merge (the one with
> the namespace changes that broke MSVC) and just going from there.
> Before that I would like comments from everyone. What would you like
> to see happen? What should not be done?
>
> _______________________________________________
> Mailing list: https://launchpad.net/~cuneiform<https://launchpad.net/%7Ecuneiform>
> Post to     : cuneiform@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~cuneiform<https://launchpad.net/%7Ecuneiform>
> More help   : https://help.launchpad.net/ListHelp
>